Changelog

Source: NEWS.md

chapr (development version)

Breaking Changes

• CLI consolidated to create_chap_cli(): The create_chapkit_cli() function has been removed. All CLI functionality is now provided by create_chap_cli(), which uses named arguments (--data, --historic, --future, --output) powered by the optparse package.

• CLI argument style changed: The CLI now uses named arguments instead of positional arguments:

  • Train: Rscript model.R train --data data.csv (was: train data.csv)
  • Predict: Rscript model.R predict --historic h.csv --future f.csv --output out.csv (was: predict h.csv f.csv model.rds)

• optparse dependency added: The CLI now uses the optparse package for argument parsing, providing better help messages and argument validation.

Enhancements

• create_chap_cli() now supports both short (-d) and long (--data) option forms
• Improved error messages for missing required arguments
• Better help text via Rscript model.R train --help

---

chapr 0.1.0

Initial CRAN Release

New Features

• Unified CLI Infrastructure: create_chap_cli() function provides single-file model development with automatic file I/O handling
  • Automatic CSV loading and tsibble conversion
  • Auto-detection of time and key columns
  • Subcommand dispatch (train/predict/info)
• Configuration Management:
  • read_model_config() for YAML configuration parsing
  • write_model_config() for configuration serialization
  • get_config_param() for safe nested parameter extraction
  • Optional schema validation with ajv package
• Model Examples:
  • Mean model baseline example demonstrating unified CLI pattern
  • EWARS spatio-temporal model example

Infrastructure

• Comprehensive test suite (28 tests)
• Complete roxygen2 documentation
• YAML configuration support (yaml package)
• Optional schema validation (ajv package)
• Proper namespacing for all external package dependencies